Middle-ear disease is the most common cause of hearing loss. Otologic surgeons repair thousands of middle ears each year, with mixed success, especially after severe middle-ear disease. Knowledge of structure-function relations is crucial to improving reconstructive techniques. In the past, we developed physiology-based models that relate structure and function in normal ears, pathological ears and reconstructed ears. We focus now on four new clinically-relevant structure- function issues: (1) We quantify the contribution of the middle ear to the shape of the auditory threshold curve by (a) comparing middle-ear inputs and outputs to quantify the power transfer efficiency of the middle-ear in individual animals, (b) comparing estimates of middle-ear output at physiologic threshold in live animals. Near constant levels of middle-ear output at threshold would imply the middle ear plays a dominant role in shaping normal auditory thresholds. (2) We investigate the contribution of the round window and natural or pathologic `third windows' to normal, pathological and aided middle-ear function. Normal and pathologic inner-ear windows are potential pathways for stimulating the cochlea as well as potential shunt paths for directing sound power away from the cochlea. This work will determine how the round window and different `third windows' into the inner ear affect sensitivity to airborne sound and assess their potential as pathways for cochlear stimulation. (3) We will quantify the relative contributions of different bone- conduction pathways. Recent reports come to contradictory conclusions on the clinical significance of compressive bone conduction, and point to a sensitivity of this mechanism to inner-ear conductive pathology. We quantify the contribution of cochlear compression relative to other bone- conduction mechanisms and test whether changes in bone-conducted thresholds with `third- window' pathologies are explained by cochlear compression. (4) We test methods, including laser- Doppler vibrometry and bone-conduction stimulation, to separate `conductive' and `sensory-neural' hearing loss phenotypes in animal models of hearing loss. This separation is critical in understanding the mechanisms of various genetic models of hearing loss.